Hi all:
How can I calculate the phonon spectrum for a molecule? We use supercell and one k point in the gS calculation. To calculate phonon spectrum we need several q points and according to the RF1 supercell is completely avoided. I can calculate the phonon frequency for q=0. can I use anaddb to get phonon spectrum?
Another question is that any one knows the example (in the tutorial) for the calculation of the Raman spectrum in Abinit 6.0.4?
Thanks
Zhifan Chen
Clark Atlanta University
Atlanta GA 30314
USA
phonon spectrum
Moderators: mverstra, joaocarloscabreu
-
pboulanger
- Posts: 3
- Joined: Fri Aug 14, 2009 4:20 pm
Re: phonon spectrum
Hi,
I don't fully understand your question, but I will give a stab at it.
Firstly, molecules are finite systems and as such they possess discrete vibrational states, i.e. there is no phonon dispersion.
Hence, you only want the discrete spectrum at the Gamma point.
As you said, to study molecules, you only need to consider a large box (to isolate one molecule from it's images). There is no need for a supercell.
You can extract the phonon frequencies for the Gamma point (Q=0.0 0.0 0.0) using the an input file similar to the one in rf1.
There is no need to use ANADDB to extract these frequencies (except if you want to calculate other properties or you want to impose the sum rule), since they are given in the output of Abinit for every Q-point.
One thing you must keep in mind, is that molecules typically have 3 translation modes, 3 (or fewer) rotation modes and finally 3N-6 vibration modes.
The three translation modes should yield zero frequencies, i.e. same as acoustic sum rule. Hence, the three first frequencies should be zero.
The rotations have a certain fundamental energy that should be small, but for certain systems, they can be hard to isolate from the other vibration frequencies.
Hope this helps,
I don't fully understand your question, but I will give a stab at it.
Firstly, molecules are finite systems and as such they possess discrete vibrational states, i.e. there is no phonon dispersion.
Hence, you only want the discrete spectrum at the Gamma point.
As you said, to study molecules, you only need to consider a large box (to isolate one molecule from it's images). There is no need for a supercell.
You can extract the phonon frequencies for the Gamma point (Q=0.0 0.0 0.0) using the an input file similar to the one in rf1.
There is no need to use ANADDB to extract these frequencies (except if you want to calculate other properties or you want to impose the sum rule), since they are given in the output of Abinit for every Q-point.
One thing you must keep in mind, is that molecules typically have 3 translation modes, 3 (or fewer) rotation modes and finally 3N-6 vibration modes.
The three translation modes should yield zero frequencies, i.e. same as acoustic sum rule. Hence, the three first frequencies should be zero.
The rotations have a certain fundamental energy that should be small, but for certain systems, they can be hard to isolate from the other vibration frequencies.
Hope this helps,